ACKNOLEDGEMENTS
The data described in this thesis comprise mainly the work performed at the Harris Birthright Research Centre for Fetal Medicine, King’s College Hospital, where I was a research fellow during my PhD project. The presence of the following people made the realization of this project possible.
I would especially like to thank Dr. Tamara Stampalija, who gave me the opportunity to have such an amazing experience and who constantly supported and encouraged me through these years, not only professionally. I’m grateful also to Prof. Giuseppe Ricci for his wise guidance and patience.
I will always be grateful to Professor Nicolaides who was a mentor, a teacher and an example in these two years. He is a constant inspiration for many doctors working in the field of fetal medicine worldwide and having had the opportunity to work with him was a precious experience.
I would also like to thank Dr. Argyro Syngelaki and Professor Ranjit Akolekar for their help and patience over these years in the collection and analysis of the data. I will always remember their constancy and they taught me that research is a combination of love and perseverance.
I will never forget the time spent with Dr. Vita Zidere and I am very grateful to her as she was instrumental in advancing my skills in echocardiography.
DECLARATION
This thesis entitled “Congenital heart defects and placental dysfunction” has been composed by me, Ilaria Fantasia, and the work on this thesis is my own. This research project was composed by me with advice from my supervisors Prof. Giuseppe Ricci and Dr. Tamara Stampalija. As part of my European PhD project, data collection and statistical analysis were held in two hospitals in UK, King’s College Hospital, under the supervision of Professor Kypros Nicolaides and Medway Maritime Hospital, under the supervision of Professor Ranjt Akolekar.
I was responsible for part of patient recruitment and acquisition of biophysical and biochemical markers at both hospitals and I participated in part of the fetal echocardiography performed under the supervision of Dr. Vita Zidere.
I was directly responsible for the collection of data and creation of the database as well as for the process of obtaining pregnancy outcomes. I wrote and composed this thesis and where information has been derived from other sources, I confirm that this has been indicated in the thesis. I contributed to writing the published paper incorporated in this thesis. This work has not previously been submitted, in part or whole, for consideration in any other degree or professional qualification.
ABBREVIATIONS
AC
Abdominal circumference
aCGH
array Comparative genomic hybridization
APS
Antiphospholipid syndrome
APVR
Abnormal pulmonary vein return
ART
Assisted reproductive techniques
CAT
Common arterial trunk
CHDs
Congenital heart defects
CNV
Copy number variant
CoA
Coarctation of the aorta
CPR
Cerebroplacental ratio
CRL
Crown-rump length
DORV
Double outlet right ventricle
DV
Ductus venosus
FGR
Fetal growth restriction
FL
Femur length
FVT
Fetal thrombotic vasculopathy
HC
Head circumference
HELLP
Haemolytic anemia, eleveted liver enzymes, low platelet
count
HLHS
Hypoplastic left heart syndrome
HRHS
Hypoplastic right sided lesions
ICSI
Intracytoplasmatic sperm injection
IQR
Interquartile range
IUGR
Intrauterine growth restriction
IVF
In-vitro fertilization
IVS
Intervilluous space
ISUOG
International Society of Ultrasound in Obstetric and
Gynecology
LSOL
Left-sided obstructive lesions
MCA
Middle cerebral artery
MoM
Multiple of the median
MRI
Magnetic resonance imaging
NDD
Neurodevelopmental delay
NT
Nuchal translucency
OR
Odds ratio
PAPP-a
Pregnancy associated plasma protein A
PE
Preeclampsia
PlGF
Placental growth factor
RSOL
Right-sided obstructive lesions
sFlt-1
Soluble Fms-like tyrosine kinase-1
SGA
Small for gestational age
SLE
Systemic lupus erythematosus
SNAs
Synticial nuclear aggregates
SVDs
Single ventricle defects
TBV
Total brain volume
TGA
Transposition of the great arteries
TOF
Tetralogy of Fallot
TOP
Termination of pregnancy
TR
Tricuspid regurgitation
UA-PI
Umbilical artery pulsatiliy index
UK
United Kingdom
US
Ultrasound
UtA-PI
uterine artery pulsatility index
VEGF
Vascular enfothelial growht factor
VSD
Ventricular septal defect
WES
Wide exome sequencing
CONTENTS
ACKNOLEDGEMENTS……… 1
DECLARATION………..2
ABBREVIATIONS………..3
LIST OF TABLES………...8
LIST OF FIGURES……….9
CHAPTER 1 INTRODUCTION………..11
1.2 Congenital heart defects: background……….11
1.2.1 Genetic, epigenetic and environmental factors………....13
1.2.2 Placental factors………..….16
1.3 Congenital heart defects and nuchal translucency………...……22
1.4 Congenital feart defects and obstetric adverse outcome………....25
1.4.1 Preeclampsia………..…..…25
1.4.2 Fetal growth and neurodevelopmental delay………..…….…27
CHAPTER 2 HYPOTHESIS………..42
2.1 Main hypothesis………...42
2.2 Specific hypothesis………...………….42
CHAPTER 3 OBJECTIVES………43
3.1 Main objective………...43
3.2 Specific objective………43
CHAPTER 4 METHODS………....44
4.1 Study population……….44
4.2 Inclusion and exclusion criteria………...
46
4.3 Statistical analysis………...
47
4.4 Classification of cardiac defects……….………….
47
CHAPTER 5 RESULTS………...49
5.1 First trimester………..
49
5.1.2 Biomarkers in outcome groups………...50
5.2 Second trimester……….
53
5.2.1 Maternal and pregnancy characteristic………...53
5.2.2 Fetal biometry and uterine arteries Doppler in all CHDs compared to control group………..………...54
5.2.3 Fetal biometry and uterine arteries Doppler in CHDs, divided by subgroups, compared to control group………..……….55
5.3 Third trimester……….
59
5.3.1 Maternal and pregnancy characteristics………59
5.3.2 Fetal biometry and fetal-maternal Dopplers in all CHDs compared to control group……….…….60
5.3.3 Fetal biometry and uterine arteries Doppler in CHDs, divided by subgroups, compared to control group……….……….62
CHAPTER 6 DISCUSSION………...69
6.1 First trimester……….69
6.1.1 Placental factors………...69 6.1.2 Environmental factors………71 6.1.3 Nuchal translucency……….……….72 6.2 Second trimester……….……….………….726.2.1 Fetal growth restriction ……….….………..72
6.2.2 Placental factors………..………..74
6.2.3 Environmental factors ……...75
6.3 Third trimester………..…….…………
76
6.3.1 Fetal biometry and fetal-maternal Dopplers………..76
6.4 Strength and limitations………...
80
6.5 Future studies………...………
80
CHAPTER 8 CONCLUSIONS………...83
CHAPTER 9 REFERENCES……….……84
LIST OF TABLES
Table 1.1 Summary of the literature.
Table 5.1 First trimester maternal and pregnancy characteristics in fetuses with congenital heart defects, stratified according to sub-groups, compared to those with normal cardiac anatomy.
Table 5.2 First trimester median and interquartile range of biomarkers in fetuses with congenital cardiac defects compared to those with a normal cardiac anatomy.
Table 5.3 First trimester correlations between biophysical and biochemical markers in fetuses with and without congenital cardiac defects.
Table 5.4 Second trimester maternal and pregnancy characteristics in fetuses with congenital cardiac defect compared to those with normal cardiac anatomy
Table 5.5 Second trimester calculated z-score for head circumference, abdominal circumference and femur length, in fetuses with congenital heart defects compared to those with normal cardiac anatomy.
Table 5.6 Second trimester median and interquartilie range of fetal biometric parameter in fetuses with congenital cardiac defects, stratified according to sub-groups, compared to those with a normal cardiac anatomy.
Table 5.7 Third trimester maternal and pregnancy characteristics in fetuses with congenital cardiac defect compared to those with normal cardiac anatomy
Table 5.8 Third trimester median and interquartilie range of biomarkers in fetuses with congenital cardiac defects compared to those with a normal cardiac anatomy
Table 5.9 Third trimester median and interquartile range of biomarkers in fetuses with congenital cardiac defects, stratified according to sub-groups compared to those with a normal cardiac anatomy.
Table 6.1 Comparisons between the results reported by Llurba et al and results of our population
LIST OF FIGURES
Figure 1.1 Locations of heart malformations that are usually identified in infancy. Figure 1.2 Embryonic development from day 2 to week 8, when development of the heart is completed.
Figure 1.3 Figure of the heart-placenta axis, showing the vascular connections between the placenta and the fetus through multiple fetal vessel that constitute the fetal circulation.
Figure 1.4 The extravillous circulations.
Figure 1.5. Box-and-whisker plot of PlGF-MoM values in the control group and in fetuses with CHD classified according to normal or abnormal NT.
Figure 1.6 A summary of the neurodevelopmental deficits observed in children who underwent corrective surgery for CHD.
Figure 1.7 Relationship between GA and TBV in fetuses with CHD and controls.
Figure 1.8 A schematic representation of the pathophysiology of neurodevelopmental deficits seen in CHD patients.
Figure 1.9 Oxygen saturations across the circulations of representative examples of a normal fetus and fetuses with hypoplastic left heart syndrome (HLHS), transposition of the great arteries (TGA) and tetralogy of Fallot (TOF) by MRI.
Figure 5.1 First trimester maternal serum placental growth factor in pregnancies with major congenital cardiac defects compared to those without defects. The cardiac defect group is subdivided according to high or normal NT and according to type of defect. Figure 5.2 Second trimester box-and-whisker plots of Z-scores for fetal head circumference, abdominal circumference, femur length and uterine artery pulsatility index in fetuses with congenital heart defects compared to those with normal cardiac anatomy.
Figure 5.4 Second trimester box-and-whisker plots of Z-scores for head circumference in fetuses with congenital heart defects, stratified according to sub-groups, compared to those with normal cardiac anatomy.
Figure 5.5 Second trimester box-and-whisker plots of Z-scores for abdominal circumference in fetuses with congenital heart defects, stratified according to sub-groups, compared to those with normal cardiac anatomy.
Figure 5.6 Second trimester box-and-whisker plots of Z-scores for femur length in fetuses with congenital heart defects, stratified according to sub-groups, compared to those with normal cardiac anatomy.
Figure 5.7 Third trimester box-and-whisker plots of Z-scores for fetal head circumference, abdominal circumference, femur length in fetuses with congenital heart defects compared to those with normal cardiac anatomy.
Figure 5.8 Third trimester box-and-whisker plots of Z-scores for umbilical artery pulsatility index, middle cerebral artery pulsatility index and uterine artery pulsatility index in fetuses with congenital heart defects compared to those with normal cardiac anatomy. Figure 5.9 Third trimester box-and-whisker plots of Z-scores for head circumference in fetuses with congenital heart defects, stratified according to sub-groups, compared to those with normal cardiac anatomy.
Figure 5.10 Third trimester box-and-whisker plots of Z-scores for abdominal circumference in fetuses with congenital heart defects, stratified according to sub-groups, compared to those with normal cardiac anatomy.
Figure 5.11 Third trimester box-and-whisker plots of Z-scores for femur length in fetuses with congenital heart defects, stratified according to sub-groups, compared to those with normal cardiac anatomy.
Fig. 5.12 Third trimester Box-and-whisker plots of Z-scores for umbilical artery pulsatility index in fetuses with congenital heart defects, stratified according to sub-groups, compared to those with normal cardiac anatomy.
Chapter 1 INTRODUCTION
1.1
CONGENITAL HEART DEFECTS: BACKGROUND
Under the name of congenital heart defects (CHD) goes a large set of structural and functional abnormalities whose origin take place in the period of embryogenesis, and that are summarized in Figure 1.1.
Figure 1.1 Locations of heart malformations that are usually identified in infancy. Numbers in brackets
CHDs are the most common type of congenital defects accounting for one third of all major congenital defects and occur in ~1% of live-born children (Matthiesen et al, 2016). CHDs represent an important medical challenge both in prenatal diagnosis, due to the required expertise in prenatal ultrasound for the correct diagnosis of the defect, and in the postnatal management, since part of these malformations requires a prompt intervention in the very first days of neonatal life.
The causes of CHD are still largely unknown. Chromosomal and single gene defects affect up to a quarter of all cases of CHD leaving the majority without an apparent explanation. Is it possible that more than half of CHDs is caused by casual “errors” during the embryogenesis or could it be that intervening external factors are able to disrupt the normal cardiac development causing the malformation?
Multi-factorial etiology, including environmental and epigenetic factors, could have a role in the pathogenesis of these “unexplained” cases, however it is still unclear how these factors interact to determine the disease. However, studying the early weeks of embryogenesis, when most of the fetal structures complete their morphogenesis, is difficult due to a lack of non-invasive techniques that allow studying the embryo in this time-frame.
1.1.1 Genetic, epigenetic and environmental factors
Un underlying parental genetic cause is known to be one of the causes of CHD: a positive family history, defined as the presence of a CHD in a first-degree relative, increases the risk for the current offspring of being diagnosed with a cardiac defect and is one of the indication for a detailed fetal echocardiography. If one of the parents is affected the risk of having a child with a cardiac anomaly is 10.7% (Huhta et al, 2013). If a previous child had a heart abnormality the recurrence risk in the subsequent pregnancy is between 1 and 4% but it is 3 to 4 times higher if two previous children were affected (Huhta et al, 2013). It is, therefore, evident how much the familiar background is determinant for the onset of CHDs, however the exact genetic risks have been difficult to identify since 90-97% of subsequent pregnancies after an affected child proceed without recurrence (Huhta et al, 2013).
The genetic cause for each specific type of cardiac abnormalities is heterogeneous and seems that genetic factors, together with epigenetic and environmental factors, are responsible for the cause of CHD. However, the way in which environmental and epigenetic factors interact with genes remains poorly understood.
Chromosomal and genetic abnormalities are a well-known risk of CHD and the incidence of chromosomal abnormalities is around 18-22% (Jansen et al, 2015). The most frequent chromosomal abnormalities are trisomy 21, trisomy 18, trisomy 13 and monosomy X.
The importance of identifying a chromosomal or a genetic defect in fetuses with CHD is related to a higher risk of associated neurodevelopment delay (NDD). However, the majority of heart defects still remains without a clear genetic or chromosomal cause in the background.
The introduction of new technologies in the genetic analysis, like single nucleotide polymorphism array, next-generation sequencing and copy number variant (CNV) platforms are widening the range of known genetic causes of cardiac malformation. This is of crucial importance when counseling the parents. A recent meta-analysis on the clinical contribution of array comparative genomic hybridization (aCGH) reported that, for isolated CHD and after karyotyping and 22q11 FISH analysis, the incremented yield was 3.4% (95% CI; 0.3-6.6%), while it was 9% for non-isolated CHD (Jansen et al, 2015). Whole exome sequencing (WES) analysis in familial cases of CHD with Mendelian inheritance without a previous known genetic cause is able to identify a likely pathogenic and pathogenic mutation in 33% of cases (LaHaye et al, 2016).
However, some isolated CHD does not follow a familial inheritance and analysis of exome sequencing in children affected by CHD found that de novo point mutations are present in several hundreds of genes that collectively contribute to 10% of severe CHD (Fahed et al, 2013). New hypothesis to explain the occurrence of de novo mutations are based on epigenetic and environmental factors that can alter the genetic background. Maternal/placental microenvironment prior to and within 5-8 weeks of conception may influence the development of the fetal organs, such as heart and central nervous system. Environmental factors that can interfere with early heart development are different and include:
- environmental teratogens (i.e. dioxin, pesticides);
- maternal exposure (alcohol, isotretinoin, thalidomide, anti-epileptic drugs); - infectious agents (i.e. rubella);
- folate deficiency;
Maternal diabetes is a well-recognized risk factor for cardiac abnormalities associated with a 4-fold increase in offspring of CHD (Oyen et al, 2016). It has been suggested that maternal hyperglycemia could alter the normal process of embryogenesis, but a population study on around 2 million births over a 34-year period, showed that improvement in perinatal care did not change significantly the rate of CHD. Therefore, other factors, like obesity, increased maternal age and pro-inflammatory state, can contribute to the strong association with heart defect in diabetic mothers. Gestational diabetes, on the opposite, is not associated to an increased risk of CHD, supporting the idea that an abnormal embryonic environment is responsible of the onset of CHD, maybe inducing epigenetic modification of CHD related genes (Oyen et al, 2016).
DNA methylation and histone modification are the most known epigenetic modifications that change chromatin regulation and thus genes expression (Chan et al, 2012). Epigenetic regulation of gene expression is one of the mechanisms involved in fetal programming. Specific gene can be activated, silenced of modulated by small non- coding RNAs (microRNA), DNA methylation status and histone modifications (Feinberg 2007).In the early embryo, DNA after fertilization undergoes progressive demethylation and becomes hypomethylated during the pluripotential stages. DNA methyltransferases are the enzymes responsible for DNA methylation. A principle source of methyl groups in the cell is S-adenosylmethionine synthetized by the folic acid metabolic cycle. Studies on mouse embryo showed that the observed cardiac defects are preventable if an adequate supplementation with folic acid is supplied early after conception and possibly at higher dose than the recommend multivitamins (Huhta et al, 2013). Therefore, epigenetic is an important area for analysis in relation to birth defects.
1.1.2 Placental factors
Early in human gestation following conception, the fertilized embryo implants in the uterine wall. The successful implantation requires adequate maternal uterine perfusion and endogenous hormone preparation to allow initial embryonic survival and later organ development and growth for a term gestation.
Shortly after implantation cardiomyocytes specification and commitment take place between days 16 and 19 post conception. The circulation and a beating tubular heart are established by 21 days post conception in human pregnancy. Next, a beating, linear, tubular heart forms that then loops and septates to form a four-chambered heart (4-6 weeks of human gestation). Human cardiac morphological development is complete by 53 days post conception (8 weeks of human gestation) (Figure 1.2).
Figure 1.2 Embryonic development from day 2 to week 8, when development of the
28 days
30 days
36 days
44 days
heart is completed (taken from http://www.embryology.ch/anglais/iperiodembry/carnegie02.html).
Initial steps in placentation occur in relatively low oxygen ambient, since in the first stages there is no maternal blood flow in the developing placenta. Histological studies on placentas of human early pregnancy (from 43 to 130 days of gestation) showed that before 8 weeks of pregnancy aggregates of cytotrophoblast cells virtually occlude the mouths of the maternal spiral arteries, ensuring that any flow into the intervillous space is a slow seepage of blood flow or even plasma filtrate. Only after this period direct channels can be observed, with increasing size and delineated shape after 11-12 weeks (Burton et al, 1999). These findings, published back in 1999, were recently confirmed by a study in which microvascular filling of the intervillous space (IVS) was demonstrated by contrast-enhanced ultrasound, giving an intravenous infusion of lipid-shelled octofluoropropane microbubbles, from 6 weeks onwards in 34 pregnant women: results showed that there is an increasing blood flow to the IVS starting from 6-7 weeks (Roberts VHJ et al, 2017).
But how the development of the embryo can happen in this under perfused environment? There are evidence that, at this stage, low levels of oxygen are essential to the normal development of the embryo for two reasons: the first is that the system is still immature to protect itself from oxidative agents; the second is that it represents a trigger for angiogenesis stimulating the production of vascular endothelial growth factor (VEGF), placental growth factor (PlGF) and angiopoietin essential to the growth of the villous and harborization of the villous tree (Charnock-Jones et al, 2000).
At later stages, physiological remodeling of the spiral arteries provides adequate blood supply increasing O2 concentration that is essential for the normal development of the
fetus: chronic states of hypoxia or sudden increased concentration of O2 can alter the
oxygen concentrations could damage the embryonic development. A normal fetal-placental homeostasis is guaranteed by O2 concentration and by the consequent
production of angiogenic factors that promote formation of the highly arborized vascular bed. There is evidence from animal studies that angiogenic factors may be implicated in cardiac morphogenesis (Llurba et al, 2013).
VEGF has many direct actions on endothelial cells, which are in some ways linked to the process of angiogenesis, and include vasodilatation, increase in micro-vascular permeability, protease release, migration and proliferation of endothelial cells and lumen foration. Studies on mice embryos have shown that it could also be involved in cardiac morphogenesis: VEGF expression is found in most endocardial cells located at point of cushion formation (Armstrong et al, 2004). In zebrafish embryos, blockage of VEGF receptors resulted in functional and structural defect of cardiac valve development, suggesting that these receptors are implicated in the formation of heart valves (Lee et al, 2006). On the basis of these findings, Lambrechts et al. performed selective genotyping on 148 families with isolated TOF and showed that the presence of specific haplotype, the AAG haplotype, which lowers VEGF expression increases the risk 1.8-fold of Tetralogy of Fallot (TOF) and is transmitted in 61% of the affected children (Lambrechts
et al, 2005). On the other side, a 2 to 3-fold overexpression of VEGF in mutant mice
embryos was also found to result in severe abnormalities of cardiac development, including an attenuated compact layer of myocardium, overproduction of trabeculae, defective ventricular septation and remodeling of the outflow tract (Miquerol et al, 2000).
Therefore, VEGF expression is not only related to specific genetic background but depends also on environmental factors, supporting the multi-factorial etiology of fetal heart defects.
PlGF, a glycoprotein belonging to the family of vascular endothelial growth factors (VEGF), is another important angiogenic factor produced by the placenta and induces proliferation, migration and activation of endothelial cells. PlGF is highly expressed by trophoblastic cells and is known to be involved in the regulation of placental vascular development.There is extensive evidence that in pregnancies with impaired placentation PlGF level, as well others angiogenic factors, is reduced and such a decrease is present from the first trimester (Tsiakkas et al, 2016).
The presence of an unbalanced angiogenic status in fetuses with a CHD is also supported by histological findings from placenta tissue in newborns with hypoplastic left heart syndrome (HLHS): they documented a reduction in the numbers of terminal villi and reduced villous vasculature (p=0.001), lower expression of PlGF RNA (p<0.05), increased in Syncityal Nuclear Aggregates (SNAs) (p<0.01) and overall reduced placental weight (p=0.02) compared to controls (Jones et al, 2015). Therefore, in fetuses with CHDs, placenta fails to expand its villous tree and to develop terminal villi. Stanek found similar findings on a group of fetuses affected by what he calls “postplacental hypoxia”: in these cases, normal perfusion is provided on maternal side but the fetus does not have enough oxygen due to the presence of specific malformation like cardiac defects, umbilical knots, etc. He found thinner and longer villi due to poor branching on the placental side as a reflection of a compromised fetal circulation (Stanek 2015). Another larger study on 120 placental histology of 120 fetuses with CHD showed that the placental weight-to-birth weight ratio was significantly reduced in CHDs than in controls and histological findings of villous hypomaturity, thrombosis, chorioangiosis and placental infarction (Rychick et al., 2018).
Linask et al., have introduced the concept of the “heart-placenta axis” (Fig. 1.3) based on the hypothesis that cardiac and placental abnormalities may coexist through polymorphisms in genetic developmental pathways common to both organs, in particular those regulated by Wnt/ß- catenin signaling, or through a lack of key micronutrients, such as folate (Linask et al. 2014).
Figure 1.3 Figure of the heart-placenta axis, showing the vascular connections between the placenta and
the fetus through multiple fetal vessel that constitute the fetal circulation: UC, umbilical cord; UA, umbilical artery; UV, umbilical vein; DV, ductus venosus; IF-OF: inflow-outflow; DA, ductus arteriosus (from Linask et al. Changes in vitelline and utero-placental hemodynamics: implications for cardiovascular development. Front. Physiol 2014. 5:390.
The heart-placental axis is associated with parallel development of the placenta and heart that utilizes many common molecules and genes and reflects intimate and synergistic growth of both organs.
However, as shown in Fig. 1.2, heart development is completed by around 8 weeks’ gestation before placental circulation has established.
Figure 1.4 The extravillous circulations. The yolk sac is the first of the extraembryonic membranes to be vascularized, and likely plays a key role in maternal-fetal transport during the period of organogenesis before the chorionic circulation is fully established at ∼12 weeks. Changes in the resistance offered by each circulation may affect gene expression and differentiation of the fetal cardiomyocytes (From Burton et al. 2018)
Extensive remodeling occurs toward the end of the first trimester when the definitive placenta is formed. Villi initially develop over the entire gestational sac but starting from around 8 weeks of gestation the villi over the superficial pole begin to regress, forming the smooth membranes or chorion laeve. Regression is associated with the progressive onset of the maternal arterial circulation to the placenta, first in the periphery and then in the rest of the placenta. This process is mediated by the migration of extravillous trophoblastic cells (EVT) into the placental bed and modulated by locally high levels of oxidative stress within the villi (Jauniaux et al., 2003).
Events at this stage of development play a key role in determining the final size and shape of the placenta, and so may impact development of the fetal heart.
1.2
Congenital heart defects and nuchal translucency
the 95th centile achieved a detection rate for CHD of 56% (Hyett et al, 1999). The prevalence of CHDs increases with increasing value of NT (3% for NT between 3.5 and 4.5 mm and 20% for NT ³5.5 mm). No differences were found according to the type of CHD, however strongest associations were seen for left-sided lesions such as HLHS and CoA. These findings are supported also by other studies that show how an increased NT thickness constitutes a risk factor for CHDs independently from the nature of the cardiac defect (Atzei et al, 2005; Syngelaki et al, 2011).
The relation between increased NT and CHDs is, however, not fully understood. Some authors advocate the presence of impaired diastolic function that leads to increased NT with a mechanism similar to that observed in severe cardiac dysfunction and fetal hydrops at later gestations, where a rise in systemic venous pressure and in hydrostatic pressure may lead to the accumulation of nuchal fluid in the first trimester (Hyett et al, 1996). In support of this theory, it is known that tricuspid regurgitation (TR) and reversed flow in the ductus venosus (DV), both signs of impaired diastolic cardiac function are more frequent in fetuses with CHDs and increased NT. However, one study examined the cardiothoracic ratio and the left ventricular ejection fraction in fetuses with HLHS and isolated ventricular septal defects and increased NT in the first trimester: they failed to prove the presence of cardiac dysfunction, though the assessment was done in the second trimester and, therefore, transient period of cardiac dysfunction that caused increased NT cannot be excluded with certainty (Simpson et al, 2000). Furthermore, additional signs of heart failure, like pericardial and pleural effusion, edema, cardiomegaly and ascites are all usually absent and increased NT can be associated to several congenital abnormalities, other than just CHDs, suggesting that there are different pathogenic pathways to the presence of a CHD and increased NT.
excluded as they were not a mutual gene involved in both cardiac and lymphatic vascular development. Consequently, 15 genes were identified as potentially mutual genes in cardiac and lymphatic vascular development. Mutations in all but one gene (Pik3ca) resulted in a cardiac defects, abnormal lymphatic development and nuchal edema. All genes were involved in the regulation of endothelial differentiation strengthening the hypothesis that abnormal endothelial differentiation, rather than cardiac failure, is the common etiologic pathway underlying both defects. No specific CHD was identified, and no specific gene was responsible for a specific CHD and this is similar to findings in clinical practice and previous studies on human fetuses with increased NT, that showed no relation between a specific cardiac defects and increased NT (Haak et al, 2005; de Mooji et al, 2010). The numerous potential interferences in this pathway explain the relative common phenotype of increased NT. However, the presence of a mutual genetic background could explain the strongest association between cardiac defects and nuchal edema, which is not marked for other fetal defects.
Figure 1.5. Box-and-whisker plot of PlGF-MoM values in the control
group and in fetuses with CHD classified according to normal or abnormal NT (From Llurba et al, 2013).
These are the first data on a possible relationship between PlGF and increased NT in fetuses with CHD however the mechanism of interaction between these two entities has not been explained. VEGF genes play an essential role in the development of lymphatic endothelial cells from veins, and mutations in some VEGF allele cause dysfunction of lymph vessels and severe systemic edema (Shibuya et
al, 2008). Since PlGF belongs to the family of VEGF the same action could be apply,
however these conclusions remain speculative since no specific study was performed so far.
1.3 Congenital heart defects and obstetric adverse outcomes
1.3.1 Preeclampsiaof pregnancy such as preeclampsia (PE) and fetal growth restriction (FGR) (Poon et
al, 2008; Akolekar et al, 2011). However, in these cases, there is an impaired
perivascular and endovascular trophoblastic invasion of the spiral arteries. As a consequence, spiral arteries fail to become low-resistance vessels, and this is reflected in increased resistance to flow in the UtA (Meekins et al, 1994).
VEGF, PlGF and s-Flt-1 are highly expressed by cytotrophoblast cells and it has been shown that their expression is altered in placenta tissue of women with preeclampsia (Zhou et al, 2002). Thus, apparently, pregnancies at risk for developing PE and pregnancies with fetus affected by a CHD share similar imbalances in the placental angiogenic environment.
Gestational hypertension didn't show significant impact in any of these cases. This study provides evidence that maternal PE and fetal CHD share a common pathway most likely linked to an endothelial dysfunction secondary to poor placental perfusion and placental insufficiency typical of earlier forms of PE, also known as “placental” forms. Unfortunately, due to the retrospective nature of the study, data on UtA-PI and maternal serum analysis of anti/angiogenic factors were not part of the analysis.
To date, there is just one retrospective study that analyzed UtA-PI in pregnancies affected by a fetal CHD in the second and third trimester: no significant differences in UtA-PI score were observed in CHD cases compared to controls but UtA-PI Z-score showed a quadratic increase with gestational age in the whole population studied and 61% of cases had UtA-PI values > 95th centile at the end of the
pregnancy. Any PE was reported in 5% of the total population studied but whether the occurrence was somehow higher in the group with increased UtA-PI was not specified (Ruiz et al, 2017).
1.3.2 Fetal growth and neurodevelopmental delay
IUGR, with an early embryonic insult responsible for the early-onset symmetric form, while the presence of placental insufficiency would be responsible for the late-onset asymmetric one (Campbell et al., 1977; Trudinger 1985; Wagner et al., 2016; Dashe
et al., 2000). Studies on fetal growth and Doppler have shown that growth restriction
secondary to placental insufficiency is characterized by increased impedance to flow in the uterine and umbilical arteries as a result of reduced placental function, while growth restriction in fetuses with fetal malformations or chromosomal abnormalities is more frequently characterized by normal Doppler values of UtA-PI and slightly higher values of UA-PI (Snijders et al., 1993; Hiersch et al., 2018).
In the Baltimore – Washington Infant Study in 1991 (Rosenthal et al, 1991) birth weight was studied according to different types of CHD and it was shown that weight at birth differs for each type of CHD and may depend on the fetal circulation determined by the defect itself. These findings were confirmed in a subsequent study by the same group (Rosenthal 1996). Fetal growth was analyzed in 4 types of CHD (TGA; TOF; HLHS; CoA) and they found that, overall, fetuses with CHD are smaller compared to controls, but the biometric parameters were different according to the type of CHD and consistent with the altered fetal circulation determined by the CHD. For example, fetuses with TGA, where deoxygenated blood is directed to the head and oxygenated blood to the body, had smaller head volume compared to the body, while fetuses with TOF, where there is a mixture of oxygenated and deoxygenated blood, were symmetrically smaller for all biometric parameters.
Around 20% of fetuses carrying a CHD are SGA and the presence of a CHD increases the risk of fetal growth impairment by two to three times (OR: 2.09) (Malik
et al, 2015). Fetal growth seems to be affected already from the second trimester
with a relative growth slope in the subsequent trimesters and HC values being the most affected (Williams et al, 2015). The relationship between CHD and fetal smallness is still unclear but it is possible that there is a shared etiologic pathway.
corrective or palliative surgery: the degree of preoperative growth failure has been associated with longer time on the ventilator, difficulties in postoperative feeding, higher risk of infection, longer hospitalization time and poor postoperative growth catch-up. Furthermore, in a recent publication, from the EPICARD study group, there are evidence that being born SGA with a major CHD requiring surgery is significantly associated with lower cognitive score than in the non-SGA group (Calderon et al, 2017). Therefore, intrauterine fetal growth is critically important in these fetuses and the presence of fetal growth restriction constitutes a negative prognostic factor especially for those cases in need of postnatal surgery.
Figure 1.6 A summary of the neurodevelopmental deficits observed in children who underwent
corrective surgery for CHD (From Nattel et al., 2017).
NDD has been usually attributed to perioperative conditions occurred during surgery resulting in cerebral hypoxia and thrombo-embolic events. However, more recently, different studies reported the presence of brain lesions at neuroimaging already before cardiac surgery as well as in cases where surgery was not performed (Khalil
et al, 2016). Magnetic resonance (MRI) studies found that the most commonly
needed to correct these defects that shouldn’t expose to serious degree of intra-operative hypoxia (Shillingford et al, 2008).
There are two main theories to explain the presence of NDD in fetuses with CHDs: the first is based on genetic and epigenetic factors that could affect the normal development of the brain in the presence of a CHD and whether any alterations of these pathways leads to abnormal development of both organs, heart and brain, with increased susceptibility of the brain tissue to hypoxic insults; the second is that the presence of a cardiac defect causes various degree of hypoxia due to the pathological fetal circulation established that secondarily affects the brain (Hinton et
al, 2008). Quite interestingly, various degrees of NDD are observed also in minor
Figure 1.7 Relationship between gestational age and Total brain
volume (TBV) in fetuses with CHD (open circles) and controls (solid diamonds) (From Limperopoulous et al. Circulation, 2010)
Figure 1.8 A schematic representation of the pathophysiology of neurodevelopmental
deficits seen in CHD patients (From Nettel et al., 2017).
FGR is most commonly caused by placental insufficiency, which exposes the fetus to a situation of chronic hypoxia. As a response to hypoxia the fetus redistributes its cardiac output to maximize oxygen and nutrient supply to the brain in a mechanism known as “brain sparing”. This happens because the fetal circulation is a parallel circuit where the majority of the right ventricular output is shunted to the descending aorta through the ductus arteriosus while the left ventricle mainly supplies the upper body and the brain. In case of placental insufficiency there is vasoconstriction of peripheral vascular beds, due to placental damage, that increases the right ventricular afterload but, on the other side, the presence of vasodilation of the cerebral arteries, due to vasodilation of the MCA, causes a decrease in the left ventricular afterload. These changes result in a preferential shift of the cardiac output in favor of the left ventricle, enhancing blood supply to the brain in what is called cerebral redistribution (Coehn et al, 2015). Changes in cerebral blood flow can be detected by measuring MCA-PI and by its ratio with the UA-PI (cerebroplacental ratio; CPR) by Doppler ultrasound.
The “brain-sparing” effect was considered a reactive mechanism in IUGR fetuses to protect the brain from hypoxia. However, there are evidence that the presence of either low MCA-PI or low CPR is significantly associated to adverse perinatal outcome and to increased risk of delay in neonatal motor and state organization, lower communication and problem–solving score at 2 years of age (Miller et a, 2016).
Figure 1.9 Oxygen saturations across the circulations of representative examples of a
normal fetus and fetuses with hypoplastic left heart syndrome (HLHS), transposition of the great arteries (TGA) and tetralogy of Fallot (TOF) by MRI. In the normal fetal circulation, there is streaming of oxygenated blood from the placenta to the fetal cerebral circulation via the ductus venosus and foramen ovale. In each of the examples of CHD, this pathway is disrupted (from Sun et al., 2016).
et al, 2005). They explain the higher MCA-PI values in RSOL to be due to
auto-regulation of the MCA to the increased blood flow directed to the brain in such lesions. The discrepancy in results with the article from D’Onofrio may be explained by the different number of cases included and the authors recommend caution in using CPR in fetuses with CHD because reference values were established in IUGR fetuses with a normal heart and they could not be of value in assessing fetuses with CHDs.
Many other studies were published on this topic but there is a high degree of heterogeneity for type of study, inclusion criteria (like gestational age, biometric parameters analyzed), classification of the heart defect adopted.
A recent meta-analysis on HC values report that in fetuses with a CHD a smaller HC is present with a value of only 0,5 SD below the population mean (Jansen et al, 2016). Data on each type of CHD could not be evaluated because of the small numbers available for each category. Fewer data are available on AC and FL values.
LVOT lesions, HLHS, TGA and TOF are the defects that were examined more frequently. MCA-PI and CPR were found to be lower in left-sided lesions, HLHS and TOF with the lowest values found in the HLHS groups (Ruiz et al, 2017; Masoller et
al, 2016; Szwast et al, 2012; Kaltman et al, 2005; Yamamoto et al, 2013; Williams et al, 2012) while contradictory results were found for TGA: some studies reported
lower values while other found similar values to controls (Ruiz et al, 2017; Jouannic
et al, 2002; Yamamoto et al, 2013; Berg et al, 2009). Fewer studies evaluated
MCA-PI in right-sided lesions, but in two that specifically compared left and right-sided lesions with healthy controls, they found the presence of a reduced MCA-PI and CPR in left-sided and increased MCA-PI in right-sided lesions (Szwast et al, 2012; Kaltman et al; 2005). Contradictory results were reported also for the UA-PI because in some reports increased values of PI were reported in fetuses with CHD, while in others there were no significant differences (Kaltman et al; 2005; Meise et al, 2001; Szwast et al).
Tab. 1.1 Summery of the literature Study (First Author, journal, Year of Publication) Study design, No. Infants
CHD Age Methods Findings
Ruiz et al, Ultrasound Obstet Gynecol, 2016 Retrospective study, N= 119
Mixed II and III
trimester Ultrasound (biometry, Doppler)
Normal MCA-PI and CPR during second trimester; 18% MCA-PI and CPR less than 5th percentile at 1st examination
Lower MCA-PI in group with severe impairment of cerebral blood flow UA-PI increased with GA Hahn et al, Ultrasound Obstet Gynecol, 2016 Retrospective study, N = 133
SVA II and III
trimester Ultrasound (biometry, Doppler)
Lower MCA-PI and decreased more as GA progressed Smaller HC at 24–29 wk GA and >34 wk GA Fetal HC predictor of neonatal HC from 30 wk GA
MCA-PI not associated with fetal and neonatal HC Zeng et al, Ultrasound Obstet Gynecol, 2015 Case-control study, N = 73/168
Mixed II and III
trimester Ultrasound (biometry, Doppler)
Lower MCA-PI
Total intracranial volume, frontal lobe volume, cerebellar volume, and thalamus volume progressively decreased from 28 wk GA
Largest decrease in frontal lobe volume, followed by total intracranial volume and cerebellar volume Smaller HC and BPD from 33
wk GA Zeng et al, Ultrasound Obstet Gynecol, 2015 Case-control study, N = 112/112 Mixed 20-30 wks Ultrasound
Higher cerebral blood flow, vascularization index, flow index, and vascularization fow index of the total intracranial volume and 3 main arteries higher in HLHS and LSOL and of the anterior cerebral artery in TGA Masoller et al, Ultrasound Obstet Gynecol, 2014 Case-control study, N = 95/95 Mixed 20-24 wks Ultrasound (biometry, Doppler)
Lower MCA-PI and CPR and higher fractional moving blood volume
Fractional moving blood volume >95th percentile in 81% compared with 11% in controls
No differences in MCA-PI and fractional moving blood volume between CHD diagnostic groups Smaller BPD and HC No differences in BPD and HC between CHD diagnostic groups Williams et al, Am Heart J, 2013 Cohort study, N = 134 SVA 18-38
wks Ultrasound (Doppler) MCA-PI at first fetal echocardiogram −0.95 ± 1.5
22% MCA-PI < −2.0 at least once across gestation Yamamoto et al, Ultrasound Obstet Gynecol, 2013 Case-control study, N = 89/89 Mixed 32 wks Ultrasound (biometry, Doppler)
Lower MCA-PI, higher UA-PI and lower CPR in HLHS and CoA
CoA with retrograde aortic arch flow, lower MCA-PI and CPR, and higher UA-PI compared with CoA with antegrade flow
Normal MCA-PI, UA-PI, and CPR in TGA and POTO Smaller HC at birth in TGA
and CoA Szwast et al, Ultrasound Obstet Gynecol, 2012 Retrospective study, N = 131/92 SVA 18-40
with pulmonary obstruction MCA-PI decreased during
gestation for aortic obstruction
MCA-PI increased during gestation for pulmonary obstruction Normal UA-PI Williams et al, Ultrasound Obstet Gynecol, 2012 Pilot study, N = 13 Mixed 20-24
wks Ultrasound (Doppler) MCA-PI −1.7 ± 1.1 56% CPR < 1.0 (no z scores) HLHS and TOF lowest MCA-PI (−2.4 and −2.01, respectively), TGA −0.75 Arduini et al, J Matern Fetal Neonatal Med, 2011 Case-control study, N = 60/65 Mixed 30-38 wks Ultrasound (biometry, Doppler)
Lower MCA-PI and CPR (no z scores)
HLHS and CoA lowest and TOF and TGA highest CPR Smaller HC and HC/AC
HLHS and CoA lowest and TOF and TGA highest HC/AC Itsukaichi et al, Fetal Diagn Ther, 2011 Retrospective study, N = 44/140 Mixed 28-34 wks Ultrasound (biometry, Doppler)
MCA-RI measurements more often less than 5th
percentile and UA-RI >90th percentile Similar biometry measurements in fetuses <10th and >10th MCA-RI percentile Berg et al, Ultrasound Obstet Gynecol, 2009 Case-control study, N = 113/137 Mixed 19-41 wks Ultrasound (biometry, Doppler)
Smaller HC at birth, normal MCA-PI and CPR in TGA Smaller HC at birth, lower
MCA-PI and CPR in HLHS Normal biometry and Doppler
parameters in PA, AoS, and TOF Guorong et al, Fetal Diagn Ther, 2009 Case-control study, N = 45/275 Mixed 20-40
wks Ultrasound (Doppler) Normal MCA-PI MCA-PI tended to be lower in LSOL and was lower in congestive heart failure Higher UA-PI and higher u/C
PI ratios
sparing” as MCA-PI was normal, whereas U/C PI was higher Kaltman et al, Ultrasound Obstet Gynecol, 2005 Case-control study, N = 58/114 Mixed 20-40
wks Ultrasound (Doppler) Lower MCA-PI in HLHS Higher MCA-PI in RSOL
compared with HLHS Higher UA-PI in RSOL Donofrio et al, Pediatr Cardiol, 2003 Case-control study, N= 36/21
Mixed II and III
trimester Ultrasound (Doppler) Lower MCA-RI and CPR Normal UA-RI
HLHS and HRHS infants had highest incidence of abnormally low CPR (58% and 60%) Jouannic et al, Ultrasound Obstet Gynecol, 2002 Case-control study, N= 23/40 TGA 36-38
wks Ultrasound (Doppler) Lower MCA-PI
Normal UA-PI, DV-PI, and Ao-PI (no z scores) Meise et al, Ultrasound Obstet Gynecol, 2001 Case-control study, N= 115/100 Mixed 19-41
wks Ultrasound (Doppler) Normal MCA-PI Higher UA-PI No difference in UA-PI >95th percentile Masoller et al, Ultrasound Obstet Gynecol, 2016 Case-control study, N = 116/116 Mixed 30-38 wks Ultrasound (biometry, Doppler)
Lower MCA-PI and CPR and higher fractional moving blood volume CHD diagnostic groups Smaller BPD and HC No differences in AC and FL between CHD and controls Modena et al, Am J Obstet Gynecol, 2006 Case-control study, N = 71/71 Mixed 24-28
wks Ultrasound (Doppler) Normal MCA-PI, UA-PI, and CPR MCA-PI more often less than 5th percentile (5/71 vs 0/71) CPR more often less than 5th percentile (8/71 vs 2/71) No difference in UA-PI >95th percentile (6/71 vs 3/71)
aorta; DV-PI, pulsatily index of the ductus venosus; HC/AC, head circumference/abdominal circumference; HRHS, hypoplastic right heart syndrome; LSOL, left-sided obstructive lesion; POTO, pulmonary outflow tract obstruction; RSOL, right-sided obstructive lesion; SVA, single ventricle anomaly; TOF, tetralogy of Fallot; U/C PI, pulsatility index of the umbilical artery/pulsatility index of the middle cerebral artery.
Overall, as found in a recently published meta-analysis (Mebius et al, 2017), it can be concluded that the existing evidence suggest a tendency towards a brain vasodilation, as reflected by MCA evaluation, mainly in those cardiac defects with impaired blood flow to the brain. However, it is not clear whether the clinical meaning of these findings is the same as in SGA fetuses.
Chapter 2 HYPOTHESIS
2.1
Main hypothesis
Isolated major CHDs are characterized by the presence of placental dysfunction.
2.2
Specific hypothesis
1. Placental dysfunction in fetuses with isolated major CHDs is reflected in an impaired angiogenic status since the first trimester of pregnancy.
2. Placental dysfunction affects fetal growth in the second and third trimester of pregnancy.
Chapter 3 OBJECTIVES
3.1 Main objective
To evaluate the correlation between isolated major CHDs and placental dysfunction and the impact on intrauterine growth and fetal Doppler throughout the pregnancy.
3.2 Specific objective
1. To evaluate the relationship between isolated major CHDs and markers of placental perfusion and function in the first trimester of the pregnancy.
2. To evaluate the relationship between isolated major CHDs, fetal growth pattern and markers of placental perfusion in the second and third trimester of the pregnancy.
Chapter 4 METHODS
4.1
Study population
The data for this study were derived from prospective screening for adverse obstetric outcomes in women attending for routine pregnancy care at King’s College Hospital and Medway Maritime Hospital, United Kingdom. The women were screened between March 2006 and October 2015 and gave written informed consent to participate in the study, which was approved by the Ethics Committee.
At the first visit, a complete recording of maternal demographic characteristics and obstetric and medical history was taken with measurement of maternal weight and height.
The first trimester ultrasound examination was performed at 11-13+6 weeks’ gestation and included: measurement of the fetal crown-rump length (CRL) to determine gestational age (Robinson et al 1975), measurement of the fetal nuchal translucency (NT) thickness (Nicolaides et al 1994), examination of the fetal anatomy for the diagnosis of major fetal defects (Syngelaki et al 2011), and transabdominal colour Doppler ultrasound for the measurement of UtA-PI (Plasencia et al 2007). The policy of both hospitals for the anatomical evaluation of the fetal heart in the first trimester is to get a color-Doppler view of the 4-chamber and the confluence of the aorta and pulmonary artery also known as “V-sign” (Syngelaki et al., 2011). Maternal serum PAPP-A and free ß-human chorionic gonadotropin were sampled for combined screening for fetal aneuploidies (Nicolaides et al 2011), while maternal serum PlGF was sampled for research purposes.
The second trimester ultrasound examination was performed at 20+0-23+6 weeks and included: estimation of fetal size from transabdominal ultrasound measurements of fetal head circumference (HC), abdominal circumference (AC) and femur length (FL) plus transabdominal color Doppler ultrasound for the measurements of the UtA-PI (Albaiges
et al 2000).
through the heart in a transverse plane to include the four-chamber view, outflow tracts and three vessel view of the heart and great vessels with and without color Doppler. If a fetal abnormality was suspected, the case was examined by a fetal medicine specialist. Likewise, all cases of suspected fetal cardiac defect were examined by a fetal cardiologist. In addition, the cardiologist carried out a fetal echocardiography at 11-14 weeks in those with NT values above the 99th centile and at 20 weeks in those with a NT between the 95th and 99th centiles.
All cases with a normal second trimester scan were offered a routine third trimester scan at 36 weeks’ gestation while those with a diagnosis of CHD were offered serials growth scan at 24+0 - 28+6 weeks, at 30+0 - 34+6 weeks, and at 35+0 - 37+6 weeks.
The ultrasound examination included, for both groups, estimation of fetal size from transabdominal ultrasound measurements of HC, AC and FL. Transabdominal colour Doppler ultrasound was used to visualize the umbilical artery (UA), the middle cerebral artery (MCA) and the UtA. Doppler parameters were recorded automatically from consecutive waveforms and measured during periods of fetal quiescence. An angle of insonation below 30° was employed. The UA waveform was recorded by assessing a single free loop of umbilical cord using the colour Doppler. The UA-PI was calculated by applying a standard formula (Acharya et al 2005). The MCA Doppler was recorded according to a standard protocol by obtaining a transverse section of the fetal head and identifying the vessel close to the circle of Willis using colour Doppler and pulsed-wave Doppler to assess impedance to flow and PI was measured when three similar consecutive waveforms were observed for all three vessels (Bahlmann et al. 2002; Vyas
et al 1990).
All neonates were examined by a pediatrician. Prenatal and neonatal findings were recorded in computerized databases. Data on pregnancy outcome from women who booked for obstetric care in our hospitals but delivered in other hospitals were obtained either from the maternity computerized records in these hospitals or the general medical practitioners of the women.
4.2 Inclusion and exclusion criteria
All cases with major cardiac defects diagnosed by pediatric cardiologists either antenatally and / or in the neonatal period were included if the subsequent criteria were met:
- measurements of maternal serum PAPP-a, bHCG, PlGF levels in the 11-13 weeks group and color Doppler UT-PI;
- measurements of fetal biometry and UtA-PI in the 20-24 weeks group;
- measurements of fetal biometry and UtA-PI, UA-PI and MCA-PI in the 30-38 weeks group.
Abnormalities suspected antenatally but not confirmed in the neonates were not included. In contrast, the prenatal diagnosis in cases of terminations and miscarriages at < 24 weeks or stillbirths at > 24 weeks were assumed to be correct because in these cases postmortem examination was not performed systematically.
The following fetal cardiac defects were not included: firstly, ventricular septal defects not requiring surgery because they are generally not considered to be major defects, secondly, right aortic arch, persistent left superior vena cava and aberrant right subclavian artery because they are not supposed to cause fetal hemodynamic changes and thirdly, cardiac tumors developing during the second and third trimesters of pregnancy because these defects would not be expected to have any manifestations during the 11-13 weeks scan.
4.3
Statistical analysis
Data from continuous variables were expressed as medians and interquartile ranges and from categorical data as n (%). Comparison of the maternal characteristics between the outcome groups was by the χ2-square test or Fisher’s exact test for categorical variables and Mann-Whitney U-test for continuous variables, respectively. A p value of < 0.05 was considered significant. Post-hoc Bonferroni correction was used for multiple comparisons.
For the first trimester the measured values of PAPP-A, PlGF and UTPI were log10 transformed to make their distributions Gaussian and each value was expressed as a multiple of the normal median (MoM) after adjustment for those characteristics that provide a substantial contribution to the log10 transformed value. The measured fetal NT was expressed as a difference from the expected normal mean for fetal CRL (delta value). Median MoM values of biomarkers were compared between outcome groups. We divided congenital cardiac defects into those with fetal NT< 3.5 and those with measurements ≥ 3.5 and compared the significant of difference in the biomarkers in each group. Non-parametric bivariate correlation analysis was used to examine the association between biomarkers in pregnancies with congenital cardiac defects and those with normal cardiac anatomy.
For the second and third trimester the observed measurements of fetal HC, AC and FL were expressed as the respective Z-score corrected for gestational age. UtA-PI was analyzed as described for the first trimester.
For the third trimester, UA-PI and MCA-PI were log10 transformed to make their distributions Gaussian and each value was expressed as a multiple of the normal median (MoM) after adjustment for those characteristics that provide a substantial contribution to the log10 transformed value.
The statistical software package SPSS 22.0 (IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp, 2013) was used for the data analyses.
4.4 Classification of cardiac defects: first and second trimester
publication from Llurba et al. (Llurba et al., 2013):
1) Conotruncal defects: tetralogy of Fallot (TOF), transposition of great arteries (TGA), double outlet right ventricle (DORV), and common arterial trunk.
2) Left ventricular outflow tract (LVOT) defects: hypoplastic left heart syndrome (HLHS), aortic stenosis (AS), coarctation of aorta (CoA), and interrupted aortic arch. 3) Valvular defects: atrioventricular septal defects (AVSD), tricuspid stenosis or atresia
CHAPTER 5
RESULTS
5.1
First trimester
5.1.1 Maternal and pregnancy characteristics
The 50,094 singleton pregnancies fulfilling the entry criteria included 49,898 pregnancies with a normal cardiac anatomy and 196 (0.4%) with major congenital cardiac defects. Of those, 73 (37.2%) were with conotruncal defects, 63 (32.1%) with LVOT defects and 60 (30.6%) with valvular abnormalities, respectively. The maternal and pregnancy characteristics are represented in Table 5.1.
Table. 5.1 Demographic and pregnancy characteristics in fetuses with congenital heart defects,
stratified according to sub-groups, compared to those with normal cardiac anatomy.
Maternal characteristics No cardiac defect (n=49,898) All cardiac defects (n=196)
Age, median (IQR) (26.7-34.8) 31.2 (26.1-36.1) 31.7
Weight, median (IQR) (59.1-77.7) 66.9 (58.5-78.9) 67.0
Height, median (IQR) (1.60-1.69) 1.65 (1.59-1.70) 1.65
SLE / APS, n (%) (0.2) 114 0
Diabetes mellitus, n (%) (0.9) 435 (2.6) * 5
Nulliparous, n (%) 25,003 (50.1) (51.5) 101
Inter-pregnancy interval, median (IQR) (2.0-4.9) 3.0 (1.8-3.9) 2.8
Post hoc Bonferroni correction for multiple comparisons; * = p< 0.0167; LVOT = left ventricular
outflow tract; IQR = interquartile range; SLE = systemic lupus erythematosus; APS = antiphospholipid syndrome.
5.1.2 Biomarkers in outcome groups
In the cardiac defect group, compared to the normal cardiac anatomy group, the median PlGF MoM and PAPP-A MoM were lower, fetal delta NT was higher and UT-PI MoM was not significantly different. In all the sub-groups of congenital cardiac defects, this trend was maintained with lower PLGF and PAPP-A MoMs, higher delta fetal NT but no significant difference in UT-PI MoM (Table 5.2, Figure 5.1).
Table 5.2. First trimester median and interquartile range of biomarkers in fetuses with congenital
cardiac defects compared to those with a normal cardiac anatomy.
Marker No cardiac defect (n=49,898) All cardiac defects (n=196) Conotruncal defects (n=73) LVOT defects (n=63) Valvular defects (n=60) Serum PAPP-A MoM 1.00 (0.69-1.42) 0.81 (0.52-1.27)** 0.73 (0.57-1.18)* 0.73 (0.44-1.29)* 0.90 (0.55-1.32) Serum PlGF MoM (0.77-1.29) 1.00 (0.56-1.07)** 0.78 (0.55-0.97)** 0.75 (0.56-1.19)* 0.80 (0.61-1.17)* 0.74 UTPI MoM (0.81-1.22) 1.00 (0.83-1.26) 1.01 (0.81-1.28) 0.97 (0.84-1.29) 1.05 (0.89-1.25) 1.00 Delta fetal NT (-0.20-0.22) 0.00 (-0.06-0.86)** 0.28 (-0.11-0.68)** 0.19 (-0.02-0.89)** 0.47 (-0.04-0.96)** 0.24
Figure 5.1 First trimester maternal serum placental growth factor in pregnancies with major
congenital cardiac defects compared to those without defects and for each group of cardiac defects. The cardiac defect group is subdivided according to high or normal NT and according to type of defect.
LVOT, left ventricular outflow tract
Table 5.3 First trimester correlations between biophysical and biochemical markers in fetuses
with and without congenital cardiac defects.
Marker Serum PlGF Congenital cardiac defects
MoM Serum PAPP-A MoM Uterine artery PI MoM Delta fetal NT Serum PlGF MoM - p<0.0001 r = 0.34; p = 0.001 r = -0.23; r = -0.15; p = 0.03 Serum PAPP-A MoM - p = 0.002 r = -0.22; r = -0.010; p = 0.9 Uterine artery PI MoM - r = -0.08; p = 0.3 Delta fetal NT -
Normal cardiac anatomy
Serum PlGF
MoM Serum PAPP-A MoM Uterine artery PI MoM Delta fetal NT Serum PlGF MoM - r = 0.29; p<0.0001 r = -0.13; p<0.0001 r = 0.004; p = 0.4 Serum PAPP-A MoM - p<0.0001 r = -0.15; p <0.0001 r = 0.02; Uterine artery PI MoM - r = -0.006; p = 0.2 Delta fetal NT -
5.2 Second trimester
5.2.1 Maternal and pregnancy characteristic
The 93,408 singleton pregnancies fulfilling the entry criteria included 92,779 pregnancies with a normal cardiac anatomy and 629 (0.7%) with major congenital cardiac defects. The maternal and pregnancy characteristics in the outcome groups are represented in Table 5.4.
Table 5.4 Second trimester maternal and pregnancy characteristics in fetuses with
congenital cardiac defect compared to those with normal cardiac anatomy
Maternal characteristics No cardiac defect (n=92,779) All cardiac defects (n=629)
Age, median (IQR) 30.9 (26.3-34.9) 31.0 (26.0-35.3)
Weight, median (IQR) 67.0 (56.9-77.6) 64.0 (59.1-72.2)
Height, median (IQR) 164.0 (160.0-168.5) 165.0 (160.0-167.74) Racial origin Caucasian, n (%) 65837 (71.0) 489 (77.7) Afro-Caribbean, n (%) 17059 (18.4) 74 (11.8) ** South Asian, n (%) 5093 (5.5) 47 (7.5) * East Asian, n (%) 2583 (2.8) 8 (1.3) Mixed, n (%) 2207 (2.4) 11 (1.7) Method of conception Spontaneous, n (%) 90593 (97.6) 611 (97.1) Assisted conception, n (%) 2186 (2.4) 18 (2.9) Cigarette smoking, n (%) 9456 (10.2) 64 (10.2) Chronic hypertension, n (%) 1207 (1.3) 1 (0.2) * SLE / APS, n (%) 179 (0.2) 1 (0.2) Diabetes mellitus, n (%) 859 (0.9) 6 (1.0) Nulliparous, n (%) 46294 (49.9) 286 (45.5)
Inter-pregnancy interval, median (IQR) 2.8 (1.8-4.7) 3.0 (2.6-3.4)
